Original Contribution

DIHYDROFLUORESCEIN DIACETATE IS SUPERIOR FOR DETECTINGINTRACELLULAR OXIDANTS: COMPARISON WITH 29,79-

DICHLORODIHYDROFLUORESCEIN DIACETATE, 5(AND 6)-CARBOXY-29,79-DICHLORODIHYDROFLUORESCEIN DIACETATE, AND

DIHYDRORHODAMINE 123

STEPHEN L. HEMPEL* ,† GARRY R. BUETTNER,† YUNXIA Q. O’MALLEY ,‡ DUANE A. WESSELS,* and

DAWN M. FLAHERTY‡

*Department of Internal Medicine, Department of Veterans Affairs Medical Center and Departments of‡Internal Medicine and†Radiology, University of Iowa, Iowa City, IA, USA

(Received21 July 1998;Revised1 March 1999;Accepted1 March 1999)

Abstract—To detect intracellular oxidant formation during reoxygenation of anoxic endothelium, the oxidant-sensingfluorescent probes, 29,79-dichlorodihydrofluorescein diacetate, dihydrorhodamine 123, or 5(and 6)-carboxy-29,79-di-chlorodihydrofluorescein diacetate were added to human umbilical vein endothelial cells during reoxygenation. None ofthese fluorescent probes were able to differentiate the controls from the reoxygenated cells in the confocal microscope.However, dihydrofluorescein diacetate demonstrated fluorescence of linear structures, consistent with mitochondria, inreoxygenated endothelium. This work tests the hypothesis that dihydrofluorescein diacetate is a better fluorescent probefor detecting intracellular oxidants because it is more reactive toward specific oxidizing species. To investigate this,dihydrofluorescein diacetate was exposed to various oxidizing species (hydrogen peroxide, superoxide [KO2], per-oxynitrite, nitric oxide, horseradish peroxidase, ferric iron, xanthine oxidase, cytochrome c, and lipoxygenase) andcompared with the three other popular probes. Though oxidized dihydrofluorescein has higher molar fluorescence,comparison of the reactions of dihydrofluorescein with these other three probes in a cell-free system indicates thatdihydrofluorescein is sometimes less fluorescent than the other probes. In addition, we find that the reactivity of all ofthe probes is very complex. Based on the results reported here, it is no longer appropriate to think of these probes asdetecting a specific oxidizing species in cells, such as H2O2, but rather as detectors of a broad range of oxidizingreactions that may be increased during intracellular oxidant stress. Cell-loading studies indicate that dihydrofluoresceinachieves higher intracellular concentrations than the second brightest intracellular probe, 29,79-dichlorodihydrofluores-cein. This fact and its higher molar fluorescence may account for the superior brightness of dihydrofluorescein diacetate.Dihydrofluorescein diacetate may be a superior fluorescent probe for many cell-based studies. © 1999 ElsevierScience Inc.

Keywords—Iron, Peroxide, Nitric oxide, Peroxynitrite, Reactive oxygen species, Free radicals

INTRODUCTION

The detection of H2O2 by 29,79-dichlorodihydrofluores-cein diacetate (DCHF-DA), (Fig. 1A), in a cell-free

system [1,2] generated interest in the use of fluorescentprobes to detect cell-derived oxidants. Shortly after pub-lication of Keston and Brandt’s papers [1,2], DCHF-DAwas used as a detector of intracellular oxidants [3]. Inthese early studies, the acetate groups were hydrolyzedwith base before adding to the cells [3]. Fluorescencewas then detected in cell lysates [3]. Cell biologists didnot use DCHF-DA to detect intracellular oxidants in livecells until 1983 when Bass demonstrated that DCHF-DAcould be used with flow cytometry to detect oxidantformation by activated neutrophils [4]. Bass speculated

Address correspondence to: Stephen L. Hempel, Department ofInternal Medicine, University of Iowa, Iowa City, IA 52242, USA; Tel:(319) 356-8265; Fax: (319) 353-6406;E-Mail: stephen-hempel@uiowa.edu.

Dr. Hempel has a primary appointment in the Department of InternalMedicine, Department of Veterans Affairs Medical Center and TheUniversity of Iowa, Iowa City, IA, and a secondary scientific appoint-ment in the Department of Radiology, The University of Iowa, IowaCity, IA.

Free Radical Biology & Medicine, Vol. 27, Nos. 1/2, pp. 146–159, 1999Copyright © 1999 Elsevier Science Inc.Printed in the USA. All rights reserved

0891-5849/99/$–see front matter

PII S0891-5849(99)00061-1

146

Fig. 1. Structures of common oxidant-sensitive fluorescent probes. (A) 29,79-dichlorodihydrofluorescein diacetate. (B) Reduced andoxidized structures of four common fluorescent probes. The nonfluorescent 29,79-dichlorodihydrofluorescein diacetate is stable inethanol and unreactive toward oxidants. After base hydrolysis of the ester bonds, releasing the acetate groups, the resulting compound,29,79-dichlorodihydrofluorescein, is reactive toward oxidizing species, such as peroxide, in the presence of appropriate cofactor(s), suchas horseradish peroxidase [2]. The nomenclature is as follows: 29,79-dichlorofluorescin diacetate or 29,79-dichlorodihydrofluorescein

147Dihydrofluorescein detects intracellular oxidants

that the probe diffuses into the cell, intracellular esteraseshydrolyze the acetate groups, and the resulting 29,79-dichlorofluorescin (DCHF) then reacts with intracellularoxidants resulting in the observed fluorescence [4]. Sincethen, there have been numerous reports using DCHF-DAto measure intracellular oxidants. Much of this work hasused flow cytometry [4–6] and cells capable of a respi-ratory burst, such as macrophages or polymorphonuclearleukocytes [4–7]. More recently, investigators have usedDCHF-DA to study other cells, such as the endothelialcell [8,9]. Additional probes have also been introduced.Dihydrorhodamine 123 (DHR123), (Fig. 1B), is pre-sumed to localize to mitochondria [10], whereas 5(and6)-carboxy-29,79-dichlorodihydrofluorescein diacetate(5&6DH-DA), (Fig. 1B), is supposed to be retained inthe cell better than DCHF, thereby yielding more fluo-rescence [11,12].

Initially, three intracellular fluorescent probes(DCHF-DA, DHR123, and 5&6DH-DA) were used inour laboratory as a tool to detect intracellular oxidantsformed during reoxygenation of anoxic endothelialcells [13–17]. We speculated that the mitochondriawould be a likely source of oxidants during reoxygen-ation [17] and attempted to use confocal microscopywith these fluorescent probes to detect mitochondria-derived species during reoxygenation. However, noneof these probes gave a fluorescent signal that wassignificantly different than the room air controls. Sub-sequently, we found that dihydrofluorescein diacetate(HFLUOR-DA) readily imaged linear intracellularstructures having the appearance of mitochondria dur-ing reoxygenation of the endothelium. These findingsindicate that HFLUOR-DA, first described as fluores-cein in 1871, may be a useful tool for other investi-gators studying intracellular oxidants [18,19].

The present study tests the hypothesis thatHFLUOR-DA is superior for detecting intracellular ox-idants by confocal microscopy because it is more reac-tive toward oxidants than the other three probes. Ourresults indicate that the superiority of HFLUOR-DA forimaging intracellular oxidants by confocal microscopy isnot due simply to enhanced reactivity. We find that thereactivity of all four probes is very complex and eachprobe has a unique reactivity toward specific oxidizingspecies. Based on results reported here, it is no longerappropriate to think of these probes as detecting a spe-cific species, such as H2O2 [6,8], but rather as detectorsof a broad range of oxidizing reactions that may beincreased during intracellular oxidant stress [20]. One

explanation for the increased fluorescence of HFLUOR-DA, compared with the other fluorescent probes, is ahigher molar fluorescence. In addition, HFLUORachieves higher intracellular concentrations than DCHF,thus more compound is available for oxidation to fluo-rescent species.

MATERIALS AND METHODS

Reagents

29,79-dichlorodihydrofluorescein diacetate (#D399),5(and 6)-carboxy-29,79-dichlorodihydrofluorescein diace-tate (#C-400), dihydrorhodamine 123 (#D632), dihy-drofluorescein diacetate (#D1194), 5(and 6)-carboxy-29,79-dichlorofluorescein (#C368), and rhodamine 123

Fig. 2. Dihydrofluorescein identifies linear structures during reoxygen-ation of anoxic endothelium. Human umbilical vein endothelial cellswere anoxic for 24 h in an atmosphere of 95% N2 and 5% CO2 at 37°C,then reoxygenated (A) for 30 min with HBSS containing 20mMfluorescein diacetate at ambient oxygen tension, 25°C. (B) Control cellswere cultured 24 h at room-air oxygen tension plus 5% CO2 thenloaded with 20mM dihydrofluorescein diacetate in HBSS. Images weregenerated by confocal microscopy using a water immersion lens. Ex-citation was 488 nm and emission was detected at 521 nm. The linearstructures visualized in the reoxygenated cells are characteristic ofmitochondria.

Fig. 1.Continued.diacetate are equivalent names of the unreactive and nonfluorescent reduced compound. 29,79-dichlorodihydrofluo-rescein or 29,79-dichlorofluorescin are equivalent names of the nonfluorescent reduced compound. 29,79-dichlorofluorescein is theoxidized fluorescent compound. The maximum excitation and emission wavelengths are shown below the structures of the oxidizedprobes. The excitation and emission wavelength will vary slightly based on pH and solvent system.

148 S. L. HEMPEL et al.

(#R302) were purchased from Molecular Probes, Inc.(Eugene, OR, USA). Powdered medium 199 (with Ear-le’s salts andL-glutamine but without NaHCO3), endo-thelial growth factor (#E2759), reduced glutathione(#G4251), glutathione peroxidase (#G6137), horseradishperoxidase, type I (#P8125), xanthine oxidase (#X1875),soybean lipoxygenase (#L6632), catalase (#C9322), Cu/Zn-SOD (#S2515), fluorescein (#F7505) and 29,79-di-chlorofluorescein (#D9053) were purchased from SigmaChemical Co. (St. Louis, MO, USA); fetal bovine serum(FBS) was purchased from Hy-Clone (Logan, UT,USA); H2O2 30% solution from Fisher Scientific (FairLawn, New Jersey), and human fibronectin from Collab-orative Research (Bedford, MA, USA). All other chem-icals were reagent grade.

Solutions

Medium 199 (M199) contained 9.87 g/l medium 199powder plus 2.2 g/l NaHCO3, 100,000 U/l penicillin G,100 mg/l streptomycin, 300 mg/lL-glutamine, plus 10%FBS, pH 7.4. HBSS contained 4.2 mM NaHCO3, pH 7.4.FeSO4 (6 mM final concentration) in 0.01 N HCl wasprepared fresh daily from 18 MV water and purged withargon for 15 min. This solution was stored by sealing itunder argon until use. In experiments that had no added

Fe21, an equal volume of 0.01 N HCl argon-purged H2Owas added to cells to serve as a control. H2O2 was mixedwith phosphate-buffered saline (PBS) (9.5 mM phos-phate), pH 7.4, before each experiment. In spite of coldstorage in the dark, there was some decay of the stock30% H2O2 solution, so the concentration was adjusted byabsorption spectroscopy according to the formula:«230 5 88 l mole21cm21. PBS was stored over Chelex-100 beads to decrease adventious iron [21]. All concen-trations quoted in the results and figure captions are finalconcentrations.

Peroxynitrite solution

Five milliliters of 0.6 M sodium nitrite in water wasadded to 5 ml 0.6 M hydrogen peroxide in acidified water(0.8 N HCl), then immediately (almost simultaneously)were added 5 ml 1.4 M sodium hydroxide in a well-stirred solution. MnO2 was then added to remove excessH2O2. The resulting solution was filtered twice through a0.2 micron Millipore (Bedford, MA, USA) filter with asyringe. The solution was then stored overnight at220°C. The peroxynitrite that rises to the top is removedand the concentration is determined from the equation«302 5 1670 M21cm21 [22].

Fig. 3. Comparison of four probes by confocal microscopy. Human umbilical vein endothelial cells were loaded with 100mMmenadione for 60 min in the presence of 20mM of different fluorescent probes in HBSS at 25°C. The cells were imaged at precisely60 min using an iris setting of 3 and a gain of 1000. The control cells were loaded with the same probe under the same conditions butwithout menadione. The microscope settings are the same for each panel such that the pixels in Panel H are saturated and little cellulardetail is apparent. At lower gain settings (not shown) more cellular detail is apparent in Panel H, as in Fig. 2. (A) DCHF, control. (B)DCHF, menadione. (C) 5&6DH, control. (D) 5&6DH, menadione. (E) DHR123, control. (F) DHR123, menadione. (G) HFLUOR,control. (H) HFLUOR, menadione.

149Dihydrofluorescein detects intracellular oxidants

Nitric oxide solution

A saturated solution of NO• was generated by thereaction of HCl with NaNO2. A 1 M solution of NaNO2

was deoxygenated by bubbling with N2 in a closed jarwith a vented headspace. After purging, a syringe wasconnected to the headspace vent. Through a separateport, small aliquots of deoxygenated 1 N HCl were addedto the NaNO2 solution with shaking. HCl injections wererepeated until no more NO• was produced. The NO• gasin the headspace was then transferred to another closedjar containing deionized water purged with N2. Theconcentration of the NO•-saturated solution in pure wateris temperature dependent (2.10 mM at 20°C; 1.93 mM at25°C).

Potassium superoxide

Reactions with KO2 were performed during condi-tions of very low relative humidity. KO2 aliquots wereweighed carefully on a Cahn 21 microanalytic balanceimmediately before use. Solutions containing all of thereagents were added with vigorous mixing in a test tubecontaining KO2. Five minutes after reacting with KO2the fluorescence was determined.

Endothelial culture

Human umbilical vein endothelial cells were isolatedfrom fresh human umbilical cords and cultured in M199with 10% FBS [16,23]. Cells were seeded on humanfibronectin-coated 100-mm dishes at a density of 43 106

cells per dish. The medium was changed after 2–4 h toremove red blood cells and nonadherent tissue cells.Previous studies have shown that this technique yieldscultures of high purity with minimal variability in cellnumber or total protein from well to well [16,24]. Cellswere passed at a 1:3 split to 35-mm dishes on day 4 and50 mg/ml endothelial cell growth factor was added. Cul-tures were examined by phase contrast microscopy priorto use to verify confluence and culture purity.

Confocal microscopy

Human umbilical vein endothelial cells were grown toconfluence in 35-mm fibronectin-coated dishes. Onehour prior to confocal imaging cells were washed withHBSS and then 20mM fluorescent probe in HBSS wasadded to the cells. A BioRad MRC-600 Confocal micro-scope with an argon-krypton laser was used to obtainimages. Cells were imaged via the epifluorescence modewith a 403 immersion lens with the following parame-ters: fluorescence transmission for fluorescein isothio-cyanate, laser power 100%, gain 865, iris 3.0, black level

0, emission filter 521. Images were stored on a MotorolaStarMax computer using the public domain NationalInstitutes of Health Image program.

Anoxia-reoxygenation

Human umbilical vein endothelial cells were culturedin a special chamber flushed with 95% N2 and 5% CO2for 24 h [16,24]. After the period of anoxia, cells werereoxygenated by washing with HBSS at ambient pO2.Then 20mM fluorescent probe was added for 30 minbefore confocal imaging.

Fluorescence spectrophotometry

In the cell-free experiments a computer-controlledPerkin-Elmer LS-50B luminescence spectrometer mea-sured changes in fluorescence of the four probes. Theunit was programmed to obtain readings at regular timeintervals, which varied depending on the experiment.Concentrations of the fluorescent probes were confirmedusing absorbance spectroscopy according to the extinc-tion coefficients provided by the supplier. DCHF-DA,HFLUOR-DA, and 5&6DH-DA were hydrolyzed withbase, 0.01 N NaOH for 30 min in the dark, then dilutedwith PBS to a working stock of 100mM. DHR123 wasdissolved in PBS and diluted to a working stock of 100mM. Reagents were added in order as depicted in theFigs. 5–15 captions. Each probe was used at a finalconcentration of 5mM for the cell-free studies. Becausethe baseline fluorescence varied from probe to probe andfrom experiment to experiment, the data presented rep-resent the change from baseline (i.e., the baseline fluo-rescence is subtracted from all values). The Exl was 488nm and the Eml was 521 nm, the same as the excitationand emission wavelengths of the confocal microscope.All fluorescence measurements were performed on roomtemperature solutions at ambient gas tension.

Cell fluorescence

Human umbilical vein endothelium at first passagewere cultured to confluence on 48 well plates. Cells werewashed with PBS, then PBS with glucose, pH 7.4, wasadded to the cells. DCHF-DA, HFLUOR-DA, DHR123,and 5&6DH-DA were added at a final concentration of20 mM. Wells without cells were also loaded with fluo-rescent probes to serve as controls. Fluorescence mea-surements were made on a Fluostar 403 (BMG LabTechnologies, Durham, NC, USA) fluorescent plate-reader every 10 min at Exl 488 nm and Eml 510(bandpass 5 nm) and 538 nm (bandpass 10 nm). The

150 S. L. HEMPEL et al.

Eml 510 nm and Eml 538 nm values were addedtogether.

Molar fluorescence

Reagent grade oxidized probes were dissolved in meth-anol to a working stock of 100mM, then diluted in 50 mMpotassium-phosphate buffer, pH 9.0, to a final concentrationof 0.5 mM. Concentrations were adjusted according to theextinction coefficients supplied by the manufacturer: 29,79-dichlorofluorescein, 91,000 cm21M21 at 502 nm; fluores-cein, 98,700 cm21M21 at 490 nm; rhodamine 123, 95,900cm21M21 at 507 nm (in methanol); and 5&6DH, 80,300cm21M21 at 504 nm.

RESULTS

Confocal microscopy

Using three popular fluorescent probes, attempts toimage mitochondria by confocal microscopy duringreoxygenation of anoxic endothelium were unsuccessful(data not shown). DCHF-DA exhibited high backgroundfluorescence in the control cells and little apparent in-crease in the reoxygenated cells. DHR123 exhibitedbright mitochondria fluorescence in the control cells aswell as the reoxygenated cells and 5&6DH-DA gave alow signal in the controls and the anoxia-reoxygenation

Fig. 4. Comparison of four probes by quantitative cell fluorescence.Human umbilical vein endothelial cells on 48 well plates were loaded with20 mM (final) of the four fluorescent probes at 25°C starting at time zeroas shown on the x axis. Fluorescence was determined by a fluorescenceplate reader, y axis. Just after the 30 min fluorescence reading, 100mMmenadione was added (arrow). Cell-free wells with the same probes andmenadione were used as negative controls. Only DHR123 in the cell-freewells had some initial fluorescence (52 arbitrary units) and slowly in-creased over 1 h (additional 19 arbitrary units, data not shown). Thesecell-free values were subtracted from the values in the figure. The oxidizedprobes depicted in the legend are 29,79-dichlorofluorescein (DCF), fluo-rescein (Fluor), rhodamine 123 (Rh 123), and 5(and 6)-carboxy-29,79-dichlorofluorescein (5 & 6);n 5 6, 6 SD.

Fig. 5.Cofactors are necessary for the oxidation of fluorescent probes byperoxide. 5mM of each probe were added to PBS, pH 7.5, and reagentsadded as described. Fluorescence was determined by a fluorescence spec-trophotometer. The excitation wavelength was set to 488 nm and thefluorescent emission frequency was monitored at 521 nm, to conform tothe wavelengths of the confocal microscope. The three probes requiringbase hydrolysis, 29,79-dichlorodihydrofluorescein diacetate, dihydrofluo-rescein diacetate, and 5(and 6)-carboxy-29,79-dichlorodihydrofluoresceindiacetate, were treated in the dark with 0.01 N NaOH for 30 min. In eachpanel,n 5 4, 6 SD. (A) At the arrows, 10mM H2O2 was added to eachof the probe solutions. (B) 30 U/ml horseradish peroxidase (HRP) wereadded, then 0.1mM H2O2 at each arrow point. Because the backgroundfluorescence at time 0 min is subtracted from each point, the absoluteintensity of Rh 123 actually exceeds the full scale (1000) of the LS50Bluminescence spectrometer. (C) 30 U/ml of HRP were added to eachprobe.

151Dihydrofluorescein detects intracellular oxidants

cells. However, dihydrofluorescein diacetate (HFLUOR-DA) readily imaged linear structures consistent withmitochondria during reoxygenation of anoxic endothe-lium (Fig. 2). To determine if HFLUOR-DA gives supe-rior results with another oxidant stress, the redox cyclingagent menadione was tested. Cells were loaded withDCHF-DA, 5&6DH-DA, DHR123, or HFLUOR-DA at

room temperature for 1 h in thepresence of menadione,then imaged by confocal microscopy. Cell fluorescenceis dependent on loading time of the fluorescent probes;therefore, these times were carefully controlled prior toobtaining the images. As shown in Fig. 3, HFLUOR-DAgives the brightest images in the control cells (panel G)and the cells treated with menadione (panel H). The

Fig. 6. Fe21 oxidizes fluorescent probes without peroxide and acts as a peroxide cofactor. (A) 20mM of Fe21, as FeSO4, was addedat 10 min. (B) 20mM H2O2 was added first then 20mM Fe21. (C) Fe31 was added, then H2O2. (D) Fe31, then 100mM ascorbate (Asc),then H2O2 were added. (E) Fe31, then 25mM EDTA, then H2O2 were added (n 5 4, 6 SD).

152 S. L. HEMPEL et al.

microscope settings are the same for each panels A–Hsuch that the pixels in panel H are saturated and littlecellular detail is apparent (compare with Fig. 2).DHR123 readily identifies linear structures consistentwith mitochondria in the control cells and the cellstreated with menadione, whereas 5&6DH-DA showsweak fluorescence. These results show important differ-ences between these probes when used for confocalmicroscopy of endothelial cells.

These observations were confirmed using a fluores-cence plate reader. The four probes were added to endo-thelial cells on 48 well plates. Additional wells withoutcells were used to measure background fluorescence.Fluorescent measurements were made every 10 min.After the 30 min measurement menadione was added(Fig. 4). These results demonstrate essentially no fluo-rescence from the cell-free wells containing DCHF-DA,HFLUOR-DA, or 5&6DH-DA (data not shown).DHR123 in the cell-free wells had some initial fluores-cence and slowly increased over 1 h. The cell-free back-ground readings were subtracted from the cell-basedfluorescence shown in Fig. 4. These results reinforce theobservations of Fig. 3, demonstrating the increased flu-orescence of HFLUOR-DA in cell-based experiments. Inaddition, the findings in the cell-free control wells con-firm that the diacetate-containing probes require interac-tion with cells to become fluorescent.

Requirement for cofactors

To test the hypothesis that dihydrofluorescein(HFLUOR) is brighter because it is more reactive towardoxidants, a series of cell-free experiments were per-formed comparing HFLUOR to the three other probes.The acetate groups were hydrolyzed with base, then the

probes were oxidized by: H2O2, H2O2 plus horseradishperoxidase (HRP), or HRP alone (Fig. 5). The spontane-ous rate of aerobic oxidation of all of the probes is low(Fig. 5A), 0–10 min. DHR123 does oxidize spontane-ously in aerated buffer, but only to 3% of full scale in 30min (data not shown). DHR123 was the most easily

Fig. 7. Glutathione peroxidase is unreactive toward the fluorescentprobes. Glutathione peroxidase, 30 U/ml, was added at 30 min. At 60min, 1 mM H2O2 was added. At 90 min, 5mM glutathione (GSH) wasadded concurrent with 1mM H2O2. At 120 min, HRP, 30 U/ml, wasadded (n 5 4, 6 SD).

Fig. 8. Catalase and Cu/Zn-SOD oxidize fluorescent probes and act ascofactors in the reaction with H2O2. (A) At the arrow, 50 U/ml catalasewere added. H2O2 10 mM is added at the arrows. (B) 5 U/ml Cu/Zn-SOD were added at the arrow, then 10mM H2O2 at the arrows. (C) 5U/ml Cu/Zn-SOD plus 50 U/mL catalase were added at the arrow, then10 mM H2O2 at the arrows (n 5 4, 6 SD).

153Dihydrofluorescein detects intracellular oxidants

oxidized by H2O2 (Fig. 5A), but oxidizes less than 10%of full scale at 50 min. In the presence of HRP all of theprobes are oxidized by H2O2; DHR123 is the most re-active, exceeding the full scale of the fluorometer (Fig.5B). Note that after several additions of H2O2 in thepresence of HRP, DCHF became unresponsive to addi-tional aliquots of H2O2 (Fig. 5B). HRP alone oxidized allof the probes, HFLUOR being only slightly more reac-tive than the others (Fig. 5C). These results demonstratethe requirement of the fluorescent probes for a specificcofactor, such as HRP [2], for optimum detection ofH2O2. The results also demonstrate that HRP alone mayoxidize the probes to fluorescence in the absence ofadded H2O2 [9,20].

A previous report suggests that Fe21 does not oxidizeDCHF to fluorescence [20]. However, Fe21 in the pres-ence of oxygen forms reactive oxidants [25]. Therefore,the oxidation of the probes in the presence of Fe21 usinga longer period of reaction than in the previous study wastested [20]. DCHF, HFLUOR, and DHR123 are oxidizedto fluorescence by Fe21, 5&6DH is not (Fig. 6A). H2O2

by itself causes little increase in fluorescence (Fig. 6B).However, the addition of Fe21 to H2O2 causes a rapidincrease in fluorescence. Again, 5&6DH has a smallincrease. The same experiment in Fig. 6B was repeatedusingtert-butylhydroperoxide. This resulted in the samerank order of reactivity, but all of the probes were lessreactive withtert-butylhydroperoxide (data not shown).Fe31 by itself, causes little increase in fluorescence andfluorescence does not increase after adding peroxide(Fig. 6C). Fe31 plus ascorbate, which reduces Fe31 toFe21, increases the oxidation of the probes, again withthe exception of 5&6DH (Fig. 6D). Fe31 plus EDTAplus H2O2 causes a modest increase in fluorescence ofDHR123 and DCHF, although there is little reaction with5&6DH or HFLUOR (Fig. 6E). These results indicate

that Fe21 plus H2O2 and Fe21 by itself oxidizes threeprobes to fluorescence; 5&6DH is the exception. Cofac-tors such as ascorbate or EDTA may enhance the reac-tivity of Fe31 with DCHF and DHR123 as previouslyreported [26]. However, HFLUOR is less reactive underthese conditions and 5&6DH is essentially not reactive.

Other intracellular cofactors

Consistent with our results, previous authors havedemonstrated that H2O2 does not readily react with thefluorescent probes without an appropriate cofactor, suchas HRP [2,9,27]. Potential cofactors in mammalian cellswere tested. Because glutathione peroxidase (GPx) ispresent in abundance in mammalian cells, GPx wastested to determine if it might participate in these kindsof reactions, either directly or by an intermediate. Figure7 shows that GPx, unlike HRP, does not activate theprobes by itself. However, the addition of H2O2 results inan increase in fluorescence of DHR123, and to a lesserextent DCHF, but not the other probes. To test if there isa glutathione intermediate that could oxidize the otherprobes, glutathione and H2O2 were added to the GPxsolution (see Fig. 7). No increase in fluorescence wasnoted. The positive control, the addition of HRP, resultedin a rapid increase in fluorescence. These results indicatethat GPx (a nonheme peroxidase) has little, if any, effecton these fluorescent probes.

Although extracellular catalase has been shown todecrease intracellular fluorescence when extracellularH2O2 is added to cells [4,8], we tested if catalase itselfmay interact directly with these probes. Surprisingly,catalase increased the fluorescence of all four probes andthe addition of H2O2 further increased the fluorescence

Fig. 9. Xanthine oxidase oxidizes the fluorescent probes. After 10 minobservation, 20 mU/ml xanthine oxidase were added to all solutions asdenoted by the arrow (n 5 4, 6 SD).

Fig. 10. Cytochrome c oxidizes dichlorodihydrofluorescein and dihy-drorhodamine 123 and acts as a peroxide cofactor. After 30 minobservation, 10mM cytochrome c was added (arrow), then 100 nMH2O2 (arrow) (n 5 4, 6 SD).

154 S. L. HEMPEL et al.

(Fig. 8A). This result indicates that catalase by itself mayact as an intracellular factor able to oxidize these probesto fluorescence [26] and that catalase may act as acofactor for the reaction with H2O2, perhaps due toperoxidase activity [28]. CuZn superoxide dismutase(Cu/Zn-SOD) can also oxidize the probes and also mayact as a cofactor for the reaction with H2O2 (Fig. 8B).Similar results are observed with catalase and Cu/Zn-SOD together (Fig. 8C). These findings demonstrate thatthese two intracellular antioxidant enzymes may them-selves oxidize the probes to fluorescence and that theseenzymes can act as cofactors for the oxidation of theseprobes by H2O2.

Zhu previously reported that xanthine oxidase couldoxidize DCHF to fluorescence [26]. These observationsare expanded in Fig. 9, demonstrating that xanthine ox-idase oxidizes all four probes, though it is most reactivetoward DHR123. The addition of xanthine (50mM) didnot alter the rate of probe oxidation (data not shown),different than reports of hypoxanthine inhibiting the ox-

idation of DCHF by xanthine oxidase [26]. These resultsindicate that xanthine oxidase, by itself, may activatethese probes to fluorescence.

Cytochrome c plus H2O2 is reported to oxidizeDHR123 [9]. The four probes were tested with cyto-chrome c and H2O2 to determine the relative reactivity(Fig. 10). These results indicate that DHR123 and DCHFincrease in fluorescence in the presence of cytochrome c.Addition of peroxide results in further increases in fluo-rescence, suggesting redox cycling (peroxidase activity)of H2O2 with cytochrome c and these two fluorescentprobes. HFLUOR and 5&6DH were not reactive withcytochrome c alone and only weakly reactive with theaddition of H2O2.

Intracellular lipid peroxides may be important mark-ers of cell oxidant stress. To determine if these probesmay detect lipid peroxides, the probes were exposed to5-lipoxygenase without and with arachidonic acid inbuffer solution (Fig. 11). Figure 11A shows that lipoxy-genase by itself will oxidize DHR123, and to a minimalextent DCHF. Arachidonic acid oxidizes none of theprobes (Fig. 11B), whereas the addition of lipoxygenaseto the arachidonic acid results in rapid oxidation ofDHR123 and less rapid oxidation of DCHF. HFLUORand 5&6DH are minimally reactive under these condi-tions. The previous results withtert-butylhydroperoxideimply that the probes will not react with organic hy-droperoxides in the absence of a cofactor, such as Fe21,or lipoxygenase.

Superoxide

Published reports indicate that the fluorescent probesDCHF and DHR123 are not oxidized by O2

•2 [8,9,26].

Fig. 11. 5-lipoxygenase plus arachidonic acid oxidizes DHR123 andDCHF. (A) At 30 min 4666 U/ml of lipoxygenase were added. (B)Arachidonic acid 150mM was added to the buffer and fluorescencemeasurements started. At 30 min 4666 U/ml of lipoxygenase wereadded.

Fig. 12. Potassium superoxide addition results in oxidization of dihy-drofluorescein and 5&6DH. KO2 was added to each of the probes at theconcentrations shown. After 5 min, a single fluorescence measurementwas recorded. The negative Rh 123 fluorescence is due to reduction ofbaseline fluorescence of Rh 123 (n 5 4, 6 SD).

155Dihydrofluorescein detects intracellular oxidants

To test this, the four probes were exposed to increasingconcentrations of KO2 (Fig. 12). KO2 decreased thefluorescence of DHR123 (there is some basal oxidationof DHR123 in solution). KO2 had little effect on DCHF,but modestly increased the fluorescence of HFLUOR and5&6DH in a dose-dependent manner.

To expand these observations and to determine ifO2

•2, in the presence of a cofactor, might oxidize theprobes to fluorescence, KO2 was added to each of thefour probes in the presence of catalase and/or Cu/Zn-SOD (Fig. 13). In Fig. 13C DHR123 fluoresced littlewith KO2 alone, but the cofactors Cu/Zn-SOD and/orcatalase increased fluorescence of DHR123 slightly.However, the other three probes fluoresced brightly withCu/Zn-SOD and/or catalase after adding KO2. HFLUORand 5&6DH were the most fluorescent, particularly in thepresence of Cu/Zn-SOD plus catalase. These results in-

dicate that KO2, in the presence of Cu/Zn-SOD and/orcatalase, oxidizes the probes to fluorescence. These find-ings imply that intracellular O2

•2 formation could oxi-dize three of the probes to fluorescence, HFLUOR and5&6DH being the most reactive. In the absence of acofactor, fluorescence is low, suggesting that O2

•2 byitself is inefficient in generating fluorescence from theseprobes.

Nitric oxide and peroxynitrite

Recent reports indicate that fluorescent probes mayreact with peroxynitrite (ONOO2) and NO• [29,30]. Fig-ure 14 demonstrates the reactivity of the probes withONOO2. It shows that DCHF is the most reactive withperoxynitrite. DHR123 increased in fluorescence afterthe first addition of ONOO2, but then became unreactive

Fig. 13. Catalase and Cu/Zn-SOD are cofactors for the oxidation of fluorescent probes by potassium superoxide. In PBS buffer 5mMfluorescent probe was exposed to buffer alone (Probe), probe plus 0.2 mg/ml KO2 (P 1 KO2), probe plus 5 U/ml Cu/Zn-SOD (P1SOD), probe plus Cu/Zn-SOD plus KO2 (P 1 SOD 1 KO2), probe plus 50 U/ml catalase (P1 Cat), probe plus catalase plus KO2

(P 1 Cat1 KO2), probe plus catalase plus Cu/Zn-SOD (P1 SOD1 Cat), and probe plus Cu/Zn-SOD plus catalase plus KO2 (P 1SOD1 Cat1 KO2). After 5 min, a single fluorescence measurement was recorded. (A) Oxidation of 29,79-dichlorodihydrofluorescein.(B) Oxidation of dihydrofluorescein. (C) Oxidation of dihydrorhodamine 123. (D) Oxidation of 5(and 6)-carboxy-29,79-dichlorodihy-drofluorescein (n 5 4, 6 SD).

156 S. L. HEMPEL et al.

to additional ONOO2. HFLUOR and 5&6DH reactedwith ONOO2 to a similar degree, but were less reactivethan DCHF. In Fig. 15, the four probes were exposed toNO•. DCHF was the only probe that fluoresced uponaddition of NO•. The relative reactivity of DCHF withONOO2 was much higher than the reactivity with NO•

(Figs. 14 and 15).

Molar fluorescence

These results examine specific reactions that mayoxidize the probes. Results are expressed as fluorescentintensity. This is a useful approach to examine specificreactions that may influence the brightness of intracellu-lar fluorescence observed under the confocal microscope.However, this approach does not account for potential

differences in the molar fluorescence of each probe.Table 1 shows the relative fluorescence of each probe atthe excitation and emission wavelengths of our confocalmicroscope and at the optimum excitation and emissionwavelengths of each probe. These results indicate thatthat fluorescein has the highest molar fluorescence of thefour probes examined.

Cell uptake of fluorescent probes

Another mechanism to account for variability of cel-lular fluorescence by different probes may be differentuptake of the fluorescent probes. To test this, cells wereloaded for 15, 30, or 60 min with DCHF-DA orHFLUOR-DA. Cells were then lysed and intracellularfluorescence measured. The nonfluorescent reduced in-tracellular probes were then oxidized with HRP andfluorescence measured again, a modification of the tech-nique previously described by Royal and Ischiropoulos[9]. Results, shown in Table 2, demonstrate that most ofthe intracellular probe is not oxidized by the cell’s basalmetabolism. After treatment with HRP, the fluorescencefrom cells treated with HFLUOR-DA is about threetimes higher than DCHF-DA–treated cells. Both probesdemonstrate substantial increases in fluorescence afterHRP treatment. This implies the presence of a largeamount of reduced probe in the cell interior that may beavailable for oxidation by an intracellular oxidant stress.The increased amount of intracellular HFLUOR avail-able for oxidation may be another factor accounting forthe higher fluorescence, as shown in Figs. 3 and 4.

DISCUSSION

This work illustrates that the oldest fluorescent probe,dihydrofluorescein [19], as the diacetate [18], is a usefulprobe for cell-based investigations. We compared themost popular probes with dihydrofluorescein diacetate,expanding observations by others, and presenting previ-ously unreported information. For example, the ability ofcatalase and Cu/Zn-SOD by themselves to increase flu-orescence of the probes is a new observation. In addition,the activity of catalase and Cu/Zn-SOD as cofactors forprobe oxidation by H2O2 is also new. During this inves-tigation several previously unobserved reactions withKO2 were also identified. It was previously reported thatthese probes are not oxidized by O2

•2 [8,9,26]. However,the results reported here demonstrate that HFLUOR and5&6DH are oxidized by the addition of KO2. The spe-cific species of reduced oxygen that results in oxidationof HFLUOR and 5&6DH when KO2 is added is un-known. Although reaction with H2O2 generated by KO2is a possibility [8,31], these probes react slowly with

Fig. 14. Peroxynitrite oxidizes the fluorescent probes. After 5 minobservation, three additions of 400 nM peroxynitrite (arrows) wereadded to 5mM fluorescent probe at 10 min intervals (n 5 4, 6 SD).

Table 1. Relative Molar Fluorescence of Oxidized Probes

Probe

Exl/Emlnm Fluorescence

Relativeto DCF

Exl/Emlnm Fluorescence

DCF 488/521 410 1.00 502/521 616Fluor 488/521 592 1.44 490/512 730Rh 123 488/521 458 1.12 499/527 5795 & 6 488/521 307 0.75 504/523 501

Reagent grade oxidized probes were used to determine the relativemolar fluorescence of each probe. 0.5mM (final) of oxidized fluores-cent probe (100mM working stock in methanol) was added to 50 mMpotassium-phosphate buffer, pH 9.0, at room temperature. The concen-trations were confirmed by the molar extinction coefficients suppliedby the manufacturer. The first set of Exl/Eml values represent theexcitation and emission wavelengths of the confocal microscope. “Rel-ative to DCF” denotes the fluorescence at Exl 488/Eml 521 comparedwith DCF. The second set of Exl/Eml values represent the optimumexcitation and emission values for the specific probes. These secondExl/Eml values were obtained by scanning and differ slightly from thepublished values shown in Fig. 1.

157Dihydrofluorescein detects intracellular oxidants

H2O2 (Fig. 5B). This implies that a species, generated byO2

•2, oxidizes HFLUOR and 5&6DH. These studies alsodemonstrate that the enzymes Cu/Zn-SOD and/or cata-lase can markedly increase the oxidation of DCHF,HFLUOR, and 5&6DH by KO2. In fact, 5&6DH, whichtended to be the least reactive probe in other reactions,was the most readily oxidized in the presence of KO2

with catalase and Cu/Zn-SOD. Finally, these resultsdemonstrate that oxidized rhodamine is reduced by KO2

(Fig. 5B), even though DHR123 is the probe mostreadily oxidized by H2O2 in our previous experiments.

The experimental results presented in Fig. 2 demon-strate the potential utility of HFLUOR in imaging sites ofcellular oxidant generation. There is a good signal-to-noise ratio, allowing HFLUOR to readily demonstratelinear fluorescence in the reoxygenated cells, whereas thecontrol cells show little fluorescence. The “brightness”observed in Fig. 3 using this probe demonstrates itsrelative superiority to the other probes under carefullycontrolled conditions. These results were confirmed us-ing a fluorescence plate reader, Fig. 4. These results

caused us to test the hypothesis that HFLUOR wasbrighter than the other probes because of enhanced re-action rate toward specific oxidants. However, the resultsin subsequent experiments did not confirm the hypothe-sis. Rather, we discovered that HFLUOR was often lessreactive toward oxidants compared to the probesDHR123 and DCHF.

Although this work did not demonstrate the specificmechanism by which HFLUOR generates brighter con-focal images, there are several potential mechanisms.One possibility is that there are additional intracellularreactions not studied here. Another possibility is that thisprobe localizes to regions of intracellular oxidant forma-tion better than the other probes. Another possibility isthat HFLUOR may be brighter due to higher molarfluorescence (see Table 1). Finally, HFLUOR-DA maybe taken up or retained by the cell better than the otherprobes as indicated by Table 2.

Though some of the reactions reported here are pre-viously described by others, this work expands much ofthe previous work by comparing the three most popularprobes to each other and introducing the oldest fluores-cent reagent [19], dihydrofluorescein, as a probe thatmay be useful for cell-based studies. These comparisonsmay assist investigators in choosing a specific probe fortheir investigations.

We have demonstrated that there are numerous reac-tions that may oxidize dihydrofluorescein and the otherthree probes. Based on these results, and previous re-ports, it is not proper to assume that these oxidant-sensing probes react only with H2O2 [6,8,20], but ratheract as detectors of a broad range of intracellular oxidiz-ing reactions.

Acknowledgments—This work was performed with funds from a De-partment of Veterans Affairs Merit Review Award, an American HeartAssociation Grant-in-Aid Award, and by a grant from the AmericanLung Association of Iowa (S.L.H.). The fluorometer measurementswere made at The University of Iowa ESR Facility, supported in partby funds from the College of Medicine, University of Iowa. Dr.Buettner and the ESR facility are supported in part by funds from theNIH, CA 66081.

Table 2. Cellular Uptake and Oxidation of DCHF-DA and HFLUOR-DA

Loading

Buffer Fluorescence Intracellular Fluorescence HRP Fluorescence

DCF Fluor DCF Fluor DCF Fluor

15 min 276 1 146 0 1156 8 1166 5 12216 153 34046 14130 min 656 5 386 4 1266 6 1246 7 19406 62 55606 11360 min 1406 5 906 3 1206 6 1216 16 20286 40 63536 342

Endothelial cells on 24 well plates were loaded with 20mM DCHF-DA or HFLUOR-DA for 15, 30, or 60 min as shown. After loading, the cellswere washed, and the extracellular buffer fluorescence (probe loading buffer plus washes) was measured at Exl 488/Eml 521 (buffer fluorescence)on the Perkin-Elmer LS-5OB luminescence spectrometer. The cells were then lysed with 0.1% SDS and the intracellular fluorescence measured(intracellular fluorescence). The cell lysates were oxidized overnight with 30 U/ml HRP (HRP fluorescence). For valid comparisons, the values forHRP fluorescence were mathematically corrected for dilution and smaller Em/Ex slit widths (n 5 4, 6 SD).

Fig. 15. Nitric oxide addition results in the oxidation of 29,79-dichlo-rodihydrofluorescein. After 4 min observation, two aliquots of 20mM(final) nitric oxide (arrows) were added to 5mM fluorescent probe (n 54, 6 SD).

158 S. L. HEMPEL et al.

REFERENCES

1. Brandt, R.; Keston, A. S. Synthesis of diacetyldichlorofluorescin: astable reagent for fluorometric analysis.Anal. Biochem.11:6–9;1965.

2. Keston, A. S.; Brandt, R. The fluorometric analysis of ultramicroquantities of hydrogen peroxide.Anal. Biochem.11:1–5; 1965.

3. Paul, B.; Sbarra, A. J. The role of the phagocyte in host-parasiteinteractions. XIII. The direct quantitative estimation of H2O2 inphagocytizing cells.Biochim. Biophys. Acta.156:168–178; 1968.

4. Bass, D. A.; Parce, J. W.; Dechatelet, L. R.; Szejda, P.; Seeds,M. C.; Thomas, M. Flow cytometric studies of oxidative productformation by neutrophils: a graded response to membrane stimu-lation. J. Immunol.130:1910–1917; 1983.

5. Rothe, G.; Valet, G. Flow cytometric analysis of respiratory burstactivity in phagocytes with hydroethidine and 29,79-dichlorofluo-rescin.J. Leukoc. Biol.47:440–448; 1990.

6. Vowells, S. J.; Sekhsaria, S.; Malech, H. L.; Shalit, M.; Fleisher,T. A. Flow cytometric analysis of the granulocyte respiratory burst:a comparison study of fluorescent probes.J. Immunol. Methods.178:89–97; 1995.

7. Hyslop, P. A.; Sklar, L. A. A quantitative fluorimetric assay for thedetermination of oxidant production by polymorphonuclear leuko-cytes: its use in the simultaneous fluorimetric assay of cellularactivation processes.Anal. Biochem.141:280–286; 1984.

8. Carter, W. O.; Narayanan, P. K.; Robinson, J. P. Intracellularhydrogen peroxide and superoxide anion detection in endothelialcells.J. Leukoc. Biol.55:253–258; 1994.

9. Royall, J. A.; Ischiropoulos, H. Evaluation of 29,79-dichlorofluo-rescin and dihydrorhodamine 123 as fluorescent probes for intra-cellular H2O2 in cultured endothelial cells.Arch. Biochem. Bio-phys.302:348–355; 1993.

10. Johnson, L. V.; Walsh, M. L.; Chen, L. B. Localization of mito-chondria in living cells with rhodamine 123.Proc. Natl. Acad. Sci.USA.77:990–994; 1980.

11. Kehrer, J. P.; Paraidathathu, T. The use of fluorescent probes toassess oxidative processes in isolated-perfused rat heart tissue.Free Radic. Res. Commun.16:217–225; 1992.

12. Morita, I.; Smith, W. L.; DeWitt, D. L.; Schindler, M. Expression-activity profiles of cells transfected with prostaglandin endoperox-ide H synthase measured by quantitative fluorescence microscopy.Biochemistry.34:7194–7199; 1995.

13. Zulueta, J. J.; Yu, F. S.; Hertig, I. A.; Thannickal, V. J.; Hassoun,P. M. Release of hydrogen peroxide in response to hypoxia-reoxygenation: role of an NAD(P)H oxidase-like enzyme in endo-thelial cell plasma membrane.Am. J. Respir. Cell Mol. Biol.12:41–49; 1995.

14. Omar, H. A.; Figueroa, R.; Omar, R. A.; Tejani, N.; Wolin, M. S.Hydrogen peroxide and reoxygenation cause prostaglandin-medi-ated contraction of human placental arteries and veins.Am. J.Obstet. Gynecol.167:201–207; 1992.

15. Lum, H.; Barr, D. A.; Shaffer, J. R.; Gordon, R. J.; Ezrin, A. M.;Malik, A. B. Reoxygenation of endothelial cells increases perme-ability by oxidant-dependent mechanisms.Circ. Res.70:991–998;1992.

16. Hempel, S. L.; Haycraft, D. L.; Hoak, J. C.; Spector, A. A.Reduced prostacyclin formation after reoxygenation of anoxicendothelium.Am. J. Physiol.259:C738–745; 1990.

17. Sanders, S. P.; Zweier, J. L.; Kuppusamy, P.; Harrison, S. J.;Bassett, D. J.; Gabrielson, E. W.; Sylvester, J. T. Hyperoxic sheeppulmonary microvascular endothelial cells generate free radicalsvia mitochondrial electron transport.J. Clin. Invest.91:46–52;1993.

18. Orndorff, W. R.; Hemmer, A. J. Fluorescein and some of itsderivatives.J. Am. Chem. Soc.49:1272–1280; 1927.

19. Baeyer.Berichte der Deutschen Chemischen Gesellschaft.4:555;1871.

20. LeBel, C. P.; Ischiropoulos, H.; Bondy, S. C. Evaluation of the

Probe 2979-Dichlorofluorescin as an Indicator of Reactive OxygenSpecies Formation and Oxidative Stress.Chem. Res. Toxicol.5:227–231; 1992.

21. Buettner, G. R. In the absence of catalytic metals ascorbate doesnot autoxidize at pH 7: ascorbate as a test for catalytic metals.J. Biochem. Biophys. Methods.16:27–40; 1988.

22. Kooy, N. W.; Royall, J. A. Agonist-induced peroxynitrite produc-tion from endothelial cells.Arch. Biochem. Biophys.310:352–359;1994.

23. Jaffe, E. A.; Nachman, R. L.; Becker, C. G.; Minick, C. R. Cultureof human endothelial cells derived from umbilical veins. Identifi-cation by morphologic and immunologic criteria.J. Clin. Invest.52:2745–2756; 1973.

24. Hempel, S. L.; Wessels, D. A.; Spector, A. A. Effect of glutathioneon endothelial prostacyclin synthesis after anoxia.Am. J. Physiol.264(6 Pt 1):C1448–1457; 1993.

25. Qian, S. Y.; Buettner, G. R. Iron and dioxygen chemistry is animportant route to initiation of biological free radical oxidations:an electroparamagnetic resonance spin trapping study.Free Radic.Biol. Med.27:1447–1456; 1999.

26. Zhu, H.; Bannenberg, G. L.; Moldeus, P.; Shertzer, H. G. Oxida-tion pathways for the intracellular probe 29,79-dichlorofluorescein.Arch. Toxicol.68:582–587; 1994.

27. Cathcart, R.; Schwiers, E.; Ames, B. N. Detection of picomolelevels of hydroperoxides using a fluorescent dichlorofluoresceinassay.Anal. Biochem.134:111–116; 1983.

28. Kalyanaraman, B.; Janzen, E. G.; Mason, R. P. Spin trapping of theazidyl radical in azide/catalase/H2O2 and various azide/peroxi-dase/H2O2 peroxidizing systems.J. Biol. Chem.260:4003–4006;1985.

29. Kooy, N. W.; Royall, J. A.; Ischiropoulos, H.; Beckman, J. S.Peroxynitrite-mediated oxidation of dihydrorhodamine 123.FreeRad. Biol. Med.16:149–156; 1994.

30. Crow, J. P. Dichlorodihydrofluorescein and dihydrorhodamine 123are sensitive indicators of peroxynitrite in vitro: implications forintracellular measurement of reactive nitrogen species.Nitric Ox-ide: Biol. Chem.1:145–157; 1997.

31. Lokesh, B. R.; Cunningham, M. L. Further studies on the forma-tion of oxygen radicals by potassium superoxide in aqueous me-dium for biochemical investigations.Toxicol. Lett. 34:75–84;1986.

ABBREVIATIONS

DCHF—29,79-dichlorodihydrofluoresceinDCHF-DA—29,79-dichlorodihydrofluorescein diacetateDHR123—dihydrorhodamine 1235&6DH—5(and 6)-carboxy-29,79-dichlorodihydrofluo-

rescein5&6DH-DA—5(and 6)-carboxy-29,79-dichlorodihydro-

fluorescein diacetateGPx—glutathione peroxidaseHFLUOR—dihdydrofluoresceinHFLUOR-DA—dihydrofluorescein diacetateHRP—horseradish peroxidaseCu/Zn-SOD—copper/zinc superoxide dismutaseDCF—29,79-dichlorofluoresceinFluor—fluoresceinRh 123—rhodamine 1235 & 6—5(and 6)-carboxy-29,79-dichlorofluorescein

159Dihydrofluorescein detects intracellular oxidants